Can AI Perform Cultural Translation? Analyzing AI-Based vs. Human Translation of Egyptian Comedy Movies as a Case Study

Audiovisual Translation (AVT) has gained widespread popularity due to technological advancements and, more importantly, audience's needs. Recently, Netflix added Modern Standard Arabic (MSA) subtitles to Egyptian Colloquial Arabic movies, which was criticized and rejected. Egyptian Arabic phrases lack bilingual terminological references and need to follow a certain methodology that reveals cultural aspects through its linguistic patterns and enables the search for equivalents in the target language. Conceptualization and perception of meaning in any language is fully loaded with cultural interpretations that are sometimes lost in translation. This paper examines subtitling Egyptian Colloquial used in movies in comparison to AI-based translation to examine whether AI-based translation can provide reliable cultural and contextual translation across languages. The research study concentrates on the dynamic nature of contextual translation, including the language that is used, and its cultural referrals. It is an interdisciplinary study designed in the spirit of critical discourse analysis with a special twist towards stylistics in text analysis, analyzing the capabilities of 5 popular AI-based tools, namely Gemini, DeepL, Perplexity, Claude, in addition to Chat GPT Plus that was trained on Egyptian Colloquial. The Egyptian movie Teer Enta is analyzed as our case study, considering five main scenes representing 5 key subtitling challenges to evaluate the generated AI-based translation compared to a publicly available online human translation. The study urges the crucial necessity for serious research involvement in cultural translation, in which AI-based translation tools are not ready yet to interpret.